Discussion
In this enriched, population-based study, we investigated the effect of diabetes status, as well as quantitative markers of glucose metabolism, on WMH volume. Confounder adjusted analysis showed a significant association between WMH volume and diabetes status. In addition, WMH volume was strongly associated with 2-hour glucose after OGTT but not with fasting glucose. HbA1c showed a significant association with WMH volume, which was attenuated after adjustment for age, gender and cardiovascular risk factors, depending on the underlying study sample.
WMH volume was significantly higher in individuals with pre-diabetes or diabetes compared with normoglycemic controls. The association between WMHs and diabetes status was confirmed after adjustment for age and gender. In the current literature, the effect of pre-diabetes on WMH volume is under debate: a recent large-scale study that defined diabetes status based on fasting glucose concentration and HbA1c levels demonstrated that severe diabetes, but not less-severe diabetes or pre-diabetes, was associated with increased WMH burden.23 On the other hand, another large-scale study that defined diabetes status based on fasting glucose concentration and 2-hour glucose concentration in OGTT reported that pre-diabetes was associated with higher WMH burden, with further deterioration in type 2 diabetes.24
We found that high 2-hour serum glucose concentration determined by OGTT, but not fasting glucose, was strongly associated with increased WMH volume. This association remained highly significant after adjustment for traditional cardiovascular risk factors. This finding might indicate that IGT, defined by OGTT, and IFG, defined by fasting glucose levels, have different effects on the development of WMHs, especially since IGT and IFG have a heterogeneous pathogenesis representing different entities in impaired glucose regulation and do not necessarily include the same group of participants.31 32 IGT is related to insulin resistance, whereas IFG is associated with impaired insulin secretion.32 Consequently, IGT and IFG may induce partially distinct pathological mechanisms with differential impact on the development of WMHs. Since pre-diabetes comprises both IGT and IFG, conflicting results of studies evaluating the association between pre-diabetes and WMHs could be partially caused by a different distribution of IGT and IFG in the underlying study samples of participants with pre-diabetes.
The stronger association of 2-hour glucose levels than fasting glucose levels or HbA1c with WMH burden might be partially explained by the known adverse cardiovascular effects of high glucose level variance and amplitude. It was demonstrated that glucose variability in OGTT was associated with greater arterial stiffness and maladaptive arterial remodeling independent of HbA1c.33 Glucose variability in continuous glucose monitoring showed an HbA1c independent association with diabetic retinopathy.34 These findings could indicate that glucose dynamics not detected by HbA1c may contribute to cardiovascular complications, potentially causing increased WMH burden.35 However, further studies are needed to assess the independent impact of glucose variability in OGTT on WMH volume.
We found no significant association between WMH burden and fasting plasma glucose across groups. In participants with established type 2 diabetes, however, a significant association between fasting glucose levels and WMHs has been reported previously.36 37 In a population-based cohort-study oversampled with type 2 diabetes, fasting glucose levels were significantly associated with WMH volume.24 In pre-diabetes however, results are more ambiguous: a recent cohort study of participants with high–normal fasting glucose levels or impaired mean fasting glucose levels (ie, pre-diabetes) found no significant difference in WMH volume between the two groups—at least when including subjects over 40 years in the model. When the analysis was restricted to participants aged 40 years and younger, the effect of mean fasting glucose levels on WMH volume was significant.38
In our study, we found a significant association of HbA1c with WMH volume, which was attenuated after adjustment for age, gender and cardiovascular risk factors, depending on the underlying study sample. This result is in keeping with the current literature: several studies have reported that elevated levels of HbA1c are associated with high WMH burden.23 39–41 Schneider et al showed in a large-scale study that severe diabetes defined by high HbA1c and long disease duration but not pre-diabetes or less-severe diabetes defined by elevated HbA1c were associated with an increased burden of WMHs.23 Tamura et al found no association of HbA1c and WMH volume in elderly patients with diabetes.35
The results of this study need to be interpreted in light of its limitations. Compared with previously published studies on WMH volume in pre-diabetes and diabetes,23 24 our sample size of 388 participants was rather small. Our results are not applicable to persons with overt cardiovascular disease, as participants with overt cardiovascular disease were excluded by study design, which was focused on MRI-based analysis of subclinical disease burden. Although traditional cardiovascular risk factors such as hypertension, obesity, dyslipidaemia, smoking and alcohol consumption were included in the analysis model, the existence of unmeasured confounding variables cannot be fully ruled out. Our results, however, indicate that high 2-hour serum glucose concentration in OGTT may be an independent risk factor for the development of WMHs, especially in pre-diabetes.
In conclusion, in this population without prior cardiovascular disease, our results suggest that the effect of diabetes and particularly pre-diabetes on WMH volume may be driven by 2-hour serum glucose concentration determined by OGTT but not by fasting glucose levels. Conflicting results of studies evaluating the association between WMHs and pre-diabetes could be partially caused by the determination of diabetes status by 2-hour glucose and fasting glucose levels plus different distribution patterns of IGT and IFG in the underlying study samples. Considering the stable association of 2-hour glucose with WMH volume, this marker could serve as a readily available parameter to inform intensified prevention strategies and to monitor early treatment strategies in individuals at risk of WMH-associated morbidity such as stroke, cognitive decline and increased risk of depression.